按索引在 pandas 列的列表中查找值

Question

我有一个包含许多列的数据框，这些列是值列表，是浮点数。我想根据索引在这些列表中找到一个或多个元素，该索引是单独列中的一个或多个整数。例如

Results                                             Index_1 Index_2

0   [2.347, 140.8, 1010.8, 723.7, 7, 0, 2.898, 9.1...   0   [0, 4, 6]
1   [93794, 112.7, 5.014, 0, 1778.1, 3473.82, 0, 3...   1   [1, 5]
2   [2.927, 12.647, 9047, 0, 1204.5, 13.4, 6.3, 4....   2   [2, 0]
3   [0, 1.801, 7.104, 2121.2, 20.375, 6.348, 11.35...   2   [2, 9, 3]

我想要创建两列，其中包含基于索引和 Index_2 的结果值。例如

Results                                             Index_1 Index_2    Outcome  Outcome_2
0   [2.347, 140.8, 1010.8, 723.7, 7, 0, 2.898, 9.1...   0   [0, 4, 6]   2.347   [2.347, 7, 2.898]
1   [93794, 112.7, 5.014, 0, 1778.1, 3473.82, 0, 3...   1   [1, 5]     112.700  [112.7, 3473.82]
2   [2.927, 12.647, 9047, 0, 1204.5, 13.4, 6.3, 4....   2   [2, 0]     9047     [9047, 2.927]
3   [0, 1.801, 7.104, 2121.2, 20.375, 6.348, 11.35...   2   [2, 9, 3]   7.104   [7.104, 12531.5, 20.375]

结果在数量上有所不同，其中一个指数也是如此。

出于某种原因，index 函数无法正常工作，我收到的错误包括：

IndexError: only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices

这是我运行df['Outcome'] = df['Results'].index['Index_1']的时候。我玩过代码，但似乎无法让它工作。

我已经检查了列的 dtype，它们 return list 用于结果，numpy.int64 用于 Index_1。

Answer 1

您可以应用可以执行此操作的 lambda 函数：

import pandas as pd
import numpy as np


df = pd.DataFrame(columns=['Results','Index_1','Index_2'],
  data=[[[2.347, 140.8, 1010.8, 723.7, 7, 0, 2.898, 9.1], 0,   [0, 4, 6]],
  [[93794, 112.7, 5.014, 0, 1778.1, 3473.82, 0, 3],1,   [1, 5]],
  [[2.927, 12.647, 9047, 0, 1204.5, 13.4, 6.3, 4],2,   [2, 0]],
  [[0, 1.801, 7.104, 2121.2, 20.375, 6.348, 11.35],2,   [2, 6, 3]]])

for x in [1,2]:
    df[f'Outcome_{x}'] = df.apply(lambda row: (np.array(row['Results'])[row[f'Index_{x}']]).tolist(), axis=1)

输出：

print(df.to_string())
                                           Results  Index_1    Index_2  Outcome_1               Outcome_2
0  [2.347, 140.8, 1010.8, 723.7, 7, 0, 2.898, 9.1]        0  [0, 4, 6]      2.347     [2.347, 7.0, 2.898]
1  [93794, 112.7, 5.014, 0, 1778.1, 3473.82, 0, 3]        1     [1, 5]    112.700        [112.7, 3473.82]
2   [2.927, 12.647, 9047, 0, 1204.5, 13.4, 6.3, 4]        2     [2, 0]   9047.000         [9047.0, 2.927]
3  [0, 1.801, 7.104, 2121.2, 20.375, 6.348, 11.35]        2  [2, 6, 3]      7.104  [7.104, 11.35, 2121.2]

按索引在 pandas 列的列表中查找值

Finding value in list in pandas column by index

python

list

python-3.x

pandas